plytrademoch

This project takes a scholarly approach to a classic market anomaly. It identifies a systematic overpricing of favorites and underpricing of longshots in certain Polymarket markets, specifically those that are 'unanchored' (no obvious external binary source of truth). The bot simulates or paper-trades a portfolio of underdog bets based on this calibrated edge. The 'measurement-first' philosophy is evident in the code structure: it includes comprehensive data collection scripts to build a historical dataset of market prices and outcomes, a statistical analysis module to quantify the bias's magnitude and confidence intervals, and then a strategy module that executes trades only when the model conditions are met. This is not a live-trading-heavy bot but a risk-management and verification framework. The code is in Python, likely using pandas and numpy for analysis, and the README details the statistical justification. The live paper-trading component sends signals via Telegram, allowing the author to monitor without risking capital. The strategy's edge is based on historical precedent (well-documented in sports and horse racing), but its transferability to Polymarket depends on market structure and participant behavior. The fork and star count are low, but the project is conceptually strong. The main risk is that the bias might disappear as more sophisticated participants enter these markets. Malware risk is low given the small, transparent codebase.